In recent years, the importance of a sensor to gather information is increasing as the IoT (Internet of Things) spreads. A data center receives information obtained by a sensor installed at a site to analyze a situation of the site and to notify the situation to a manager.
It is known to use a camera as a sensor. It is possible to confirm whether any abnormality occurs or not at the site by performing image analysis on the image which is imaged by the onsite camera. For instance, it is known a method for detecting the existence of any suspicious person at the place from the image and informing the detection to the manager.
When applying the method of identifying information by analyzing the image, it is difficult to identify information in case there is no remarkable difference in the feature amount of information to be identified compared with a usual image. For instance, when a suspicious person shows up within the image range of the camera as described above, it is not difficult to detect a suspicious person in case the person is imaged with a camera by which no person is usually imaged, but it is difficult to identify a suspicious person among the persons in case the person is imaged with a camera by which many persons are usually imaged. That is, the way to calculate the feature amount to identify information becomes a problem.
On the other hand, a method of identifying information to analyze the feature amount of an image of person's face is disclosed (for instance, see Patent Document 1). In Patent Document 1, a reliability of the feature amount data based on the size of the face area is calculated from the image data and matched to the database to recognize the object person. Thus, it is known that information can be identified by extracting the feature amount of the image and matching the database concerning the feature amount.